CN115190217A - Data security encryption method and device fusing self-coding network - Google Patents
Data security encryption method and device fusing self-coding network Download PDFInfo
- Publication number
- CN115190217A CN115190217A CN202210801788.0A CN202210801788A CN115190217A CN 115190217 A CN115190217 A CN 115190217A CN 202210801788 A CN202210801788 A CN 202210801788A CN 115190217 A CN115190217 A CN 115190217A
- Authority
- CN
- China
- Prior art keywords
- picture
- data
- result
- convolution
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000007906 compression Methods 0.000 claims abstract description 45
- 230000006835 compression Effects 0.000 claims abstract description 43
- 230000009466 transformation Effects 0.000 claims description 69
- 238000000605 extraction Methods 0.000 claims description 21
- 239000011159 matrix material Substances 0.000 claims description 17
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 238000003672 processing method Methods 0.000 claims description 3
- 230000005540 biological transmission Effects 0.000 abstract description 15
- 238000012545 processing Methods 0.000 abstract description 14
- 238000013500 data storage Methods 0.000 abstract description 4
- 230000007547 defect Effects 0.000 abstract description 2
- 238000012886 linear function Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 101150116295 CAT2 gene Proteins 0.000 description 4
- 101100326920 Caenorhabditis elegans ctl-1 gene Proteins 0.000 description 4
- 101100494773 Caenorhabditis elegans ctl-2 gene Proteins 0.000 description 4
- 101100112369 Fasciola hepatica Cat-1 gene Proteins 0.000 description 4
- 101100005271 Neurospora crassa (strain ATCC 24698 / 74-OR23-1A / CBS 708.71 / DSM 1257 / FGSC 987) cat-1 gene Proteins 0.000 description 4
- 101100126846 Neurospora crassa (strain ATCC 24698 / 74-OR23-1A / CBS 708.71 / DSM 1257 / FGSC 987) katG gene Proteins 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/44—Secrecy systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/602—Providing cryptographic facilities or services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
- G06T9/002—Image coding using neural networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/04—Network architectures or network communication protocols for network security for providing a confidential data exchange among entities communicating through data packet networks
- H04L63/0428—Network architectures or network communication protocols for network security for providing a confidential data exchange among entities communicating through data packet networks wherein the data content is protected, e.g. by encrypting or encapsulating the payload
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computer Security & Cryptography (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Bioethics (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Facsimile Transmission Control (AREA)
Abstract
The invention discloses a data security encryption method and device fusing a self-coding network, and relates to the technical field of internet data processing. The invention aims to solve the defects of high cost of mass data storage resources containing a large number of pictures, unsafe data transmission and low data transmission efficiency in the process of safely encrypting the existing data, and the method comprises the steps of encrypting the text type data by adopting a text encryption module, constructing a picture self-coding network model, and pre-compressing the original picture type data to be encrypted by adopting a picture compression module; the image compression coding is encrypted by adopting an image encryption module, the text ciphertext data or the image ciphertext data which are required to be applied to downstream tasks are decrypted by adopting a decryption module, the decrypted image compression coding is reconstructed and restored by adopting an image reconstruction module, and the decoder reconstructs the code words to obtain reconstructed image type data. The invention is mainly used for mass data transmission.
Description
Technical Field
The invention relates to the technical field of internet data processing, in particular to a data security encryption method and device.
Background
In recent years, with rapid development of information technologies such as cloud computing, big data, internet of things and the like and traditional industrial digital transformation, data sources and quantity are increasing at an unprecedented rate. Data is a new production factor, and the safety of the data is also an important research subject. Encryption processing of data in data storage and transmission processes is one of effective means for ensuring data security. However, the existing data encryption technology does not distinguish the type of data to be encrypted, the same data processing mode and encryption mode are adopted for both picture data and text data, a large amount of encrypted data can be generated during transmission of picture encryption, and when the existing data encryption technology is oriented to mass data containing a large amount of pictures, the problems of high cost of encrypted ciphertext data storage resources, low transmission efficiency and the like exist.
Therefore, a data security encryption method and device for a converged self-coding network, which can handle mass data including a large number of pictures, is safe in data transmission and high in transmission efficiency, is needed.
Disclosure of Invention
The invention aims to overcome the defects of high cost of mass data storage resources containing a large number of pictures, unsafe data transmission and low data transmission efficiency in the existing data security encryption process, and provides a data security encryption method and device which can deal with mass data containing a large number of pictures, and is integrated with a self-coding network and has safe data transmission and high transmission efficiency.
The invention relates to a data security encryption method fusing a self-coding network, which comprises the following steps:
s1, dividing all data to be processed into original text type data and original picture type data;
s2, encrypting original text type data in the data to be encrypted by adopting a text encryption module to obtain text ciphertext data;
s3, constructing a picture self-coding network model, and pre-compressing original picture type data to be encrypted by adopting a picture compression module to obtain picture compression codes; the picture self-coding network model comprises an encoder and a decoder, wherein the encoder compresses and encodes picture type data to form code words;
s4, encrypting the picture compression codes by adopting a picture encryption module to obtain picture ciphertext data;
s5, transmitting or storing the text ciphertext data and the picture ciphertext data according to user requirements;
s6, decrypting the text ciphertext data or the picture ciphertext data which need to be applied to the downstream task by adopting a decryption module to obtain decrypted text type data and decrypted picture compression codes;
and S7, reconstructing and restoring the decrypted picture compression codes by adopting a picture reconstruction module, and reconstructing the code words by the decoder to obtain reconstructed picture type data.
Further, the method comprises the following steps: in S2 and S4, the encryption processing methods adopted by the text encryption module and the picture encryption module include a symmetric encryption algorithm and/or an asymmetric encryption algorithm.
Further, the method comprises the following steps: in S2, in S3, the encoder portion includes an input layer, a convolutional layer, and a feature extraction layer, and the decoder includes a convolutional layer, a feature extraction layer, and an output layer.
Further, the method comprises the following steps: in S3, the pre-compression process adopted by the picture compression module includes the following steps:
s31, taking the original picture type data as the input of the encoder part, and defining the depth, the width and the height of the original picture type data;
s32, passing the defined picture type data through a convolution layer, performing linear transformation and coding twice on the result after convolution, and splicing the result after the linear transformation twice and the result after the linear transformation once to obtain a convolution splicing result;
and S33, carrying out secondary convolution on the primary convolution splicing result through a convolution layer to enable the depth of the primary convolution splicing result to be unchanged, splicing the result after the secondary convolution and the result of the primary convolution and carrying out linear transformation to obtain a code word with the unchanged depth and the width and the height reduced to 1/8 of the original picture type data.
Further: in S7, the reconstruction processing of the picture reconstruction module includes the following steps:
s71, inputting the code words into a convolution layer after linear transformation and matrix transformation to carry out primary convolution to obtain a primary convolution result;
s72, after the characteristics of the primary convolution result are extracted, linear transformation and matrix transformation are carried out again to obtain a primary transformation result;
s73, after feature extraction is carried out on the primary convolution result twice, linear transformation and matrix transformation are carried out again to obtain a secondary transformation result;
s74, performing three-time feature extraction on the primary convolution result, and performing linear transformation and matrix transformation again to obtain a three-time transformation result;
s75, splicing the secondary linear transformation result and the tertiary linear transformation result, and performing convolution and twice convolution on the spliced result to obtain a secondary convolution result;
and S76, splicing the primary convolution result and the secondary convolution result, and convolving the spliced result to obtain a reconstructed picture with the same dimensionality as the original picture type data.
Further, the method comprises the following steps: in S7, the cost function of the picture self-coding network model is a mean square error of the reconstructed picture and the original picture type data.
Further: in S2, the feature extraction layer includes a dynamic position code, a multi-head relation aggregator, and a feed-forward neural network.
The invention relates to a data security encryption device, which comprises a text encryption module, a picture compression module, a picture encryption module, a decryption module and a picture reconstruction module;
the text encryption module is used for encrypting the original text type data so as to obtain text ciphertext data;
the picture compression module is used for pre-compressing the original picture type data so as to obtain picture compression codes;
the picture encryption module is used for encrypting the picture compression codes so as to obtain picture ciphertext data;
the decryption module is used for decrypting the text ciphertext data or the picture ciphertext data which need to be applied to a downstream task according to user requirements to obtain decrypted text type data and decrypted picture compression codes;
the picture reconstruction module is used for reconstructing and restoring the picture compression codes so as to obtain reconstructed picture type data.
The invention has the beneficial effects that:
the data security encryption method fusing the self-coding network, which is disclosed by the invention, aims at the characteristics of massive internet data scale, various types and the like, introduces the self-coding network in the data security encryption process, realizes the data security encryption, and simultaneously reduces the bandwidth resource and hardware equipment requirements in the storage and transmission processes. A self-coding model UniFormer-based automatic Encoder (self-coding network model) including a coder and a decoder structure is built on the basis of a UniFormer (Unified Transformer) network. In the process of encrypting the picture type data to be encrypted, compression coding preprocessing is added, the picture type data is compressed and coded into low-dimensional code words by utilizing a model coder part in advance, original picture type data is reconstructed by utilizing a model decoder part in a decryption stage, the picture type ciphertext data is used for decryption and reconstruction restoration before subsequent downstream tasks, the reconstruction result is nearly lossless, and resources and bandwidth expenses required by storage and transmission of the data can be reduced while the data safety is guaranteed.
Drawings
FIG. 1 is a functional block diagram of an Encoder Encoder;
FIG. 2 is a functional block diagram of a Decoder;
FIG. 3 is a schematic diagram of a coding unit;
FIG. 4 is a schematic diagram of the structure of an encoder;
fig. 5 is a functional block diagram of a data security encryption apparatus.
Detailed Description
The following are only preferred embodiments of the present invention, but the scope of the present invention is not limited thereto, and any modifications or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. The following examples are only for illustrating the present invention and should not be construed as limiting the present invention, which should be subject to the protection scope of the claims. Embodiments of the present invention are described in detail below, and for convenience of describing the present invention and simplifying the description, technical terms used in the description of the present invention should be interpreted broadly, including but not limited to conventional alternatives not mentioned in the present application, including both direct and indirect implementations.
Example 1
The embodiment is described with reference to fig. 1 to fig. 5, and the data security encryption method disclosed in the embodiment includes the following steps:
s1, dividing all data to be processed into original text type data and original picture type data;
s2, encrypting original text type data in data to be encrypted by adopting a text encryption module to obtain text ciphertext data;
s3, constructing a picture self-coding network model, and pre-compressing original picture type data to be encrypted by adopting a picture compression module to obtain picture compression codes; the picture self-coding network model comprises an encoder and a decoder, wherein the encoder compresses and encodes picture type data to form a code word; and coding the picture type data in the data to be encrypted by utilizing the coder part of the trained UniFormer-based AutoEncoder model to obtain the code word information after the picture type data is compressed and coded.
S4, encrypting the picture compression codes by adopting a picture encryption module to obtain picture ciphertext data; carrying out encryption processing on the processed picture compression codes;
s5, transmitting or storing the text ciphertext data and the picture ciphertext data according to user requirements;
s6, decrypting the text ciphertext data or the picture ciphertext data which need to be applied to the downstream task by adopting a decryption module to obtain decrypted text type data and decrypted picture compression codes; the original text type data and the coded data after the picture type text processing can be obtained through the decryption module. The encryption and decryption methods used are all existing algorithms, and if the S2 and S4 adopt the asymmetric encryption algorithm RSA, the algorithm is also used for decryption. The RSA algorithm generates a pair of matched public key and private key, and uses the public key to encrypt data and the private key to decrypt the data.
And S7, reconstructing and restoring the decrypted picture compression codes by adopting a picture reconstruction module, and reconstructing the code words by the decoder to obtain reconstructed picture type data.
And reconstructing and restoring the compressed picture by utilizing the constructed picture self-coding network. And calling a decoder part of the trained UniFormer-based AutoEncoder model, and coding the decrypted picture type data as input to obtain the reconstructed picture type data. And at this point, the decryption of the ciphertext data is completed.
Example 2
In the embodiment, the embodiment is described with reference to embodiment 1, and in S2 and S4, the encryption processing methods adopted by the text encryption module and the picture encryption module include a symmetric encryption algorithm and/or an asymmetric encryption algorithm. And encrypting the text type data in the data to be encrypted, wherein the encryption algorithm comprises but is not limited to a symmetric encryption algorithm such as AES (advanced encryption standard) or an asymmetric encryption algorithm such as RSA (rivest-Shamir-Adleman).
Example 3
In the data security encryption method disclosed in this embodiment, in S3, the encoder portion includes an input layer, a convolutional layer, and a feature extraction layer, and the decoder includes the convolutional layer, the feature extraction layer, and an output layer.
Example 4
In the embodiment, the present embodiment is described with reference to embodiment 1, and in S3, the pre-compression processing adopted by the picture compression module includes the following steps:
s31, taking the original picture type data as the input of the encoder part, and defining the depth, the width and the height of the original picture type data;
s32, passing the defined picture type data through a convolution layer, performing linear transformation and coding twice on the result after convolution, and splicing the result after the linear transformation twice and the result after the linear transformation once to obtain a convolution splicing result;
and S33, carrying out secondary convolution on the primary convolution splicing result through a convolution layer to enable the depth of the primary convolution splicing result to be unchanged, splicing the result after the secondary convolution and the result of the primary convolution and carrying out linear transformation to obtain a code word with the unchanged depth and the width and the height reduced to 1/8 of the original picture type data.
Inputting the picture type data into a convolution layer after linear transformation and matrix transformation to carry out primary convolution to obtain a primary convolution result; after the characteristics of the primary convolution result are extracted, linear transformation and matrix transformation are carried out again to obtain a primary transformation result;
performing feature extraction twice on the primary convolution result, and performing linear transformation and matrix transformation again to obtain a secondary transformation result; performing three times of feature extraction on the primary convolution result, and performing linear transformation and matrix transformation again to obtain a three times of transformation result;
splicing the secondary linear transformation result and the tertiary linear transformation result, and performing convolution and twice convolution on the spliced result to obtain a secondary convolution result;
and splicing the primary convolution result and the secondary convolution result, and performing convolution and linear transformation on the spliced result to obtain a code word of the picture type data.
A model UniFormer-based AutoEnder (Uniform Transformer-based self-Encoder) including Encoder (Encoder) and Decoder (Decode) structures was constructed and trained. Taking the picture type data p as an example, let p be c × h × w in size, i.e. the number of channels of picture p is c, the height is h, and the width is w.
Taking the picture p as the input of a coder part of a UniFormer-based automatic encoder (self-coding network), firstly passing through a convolutional layer, inputting the result obtained by the convolutional layer into a feature extraction layer (a first UniFormer unit), and recording the output result as s0. Calling a Linear function (Linear transformation function) for s0, and transforming a processing result reshape (matrix transformation) to c × h' × w ″ (h ″)<h，w`<Such as
) The size is denoted as s _0. S0 is input to the feature extraction layer (second uniform unit), and the output result is denoted as s1. For s1, a Linear function is called, and the processing result is similarly reshape with a size of c × h '× w', which is denoted as s _1.
S1 is input into the feature extraction layer (third uniform unit), and the output result is denoted as s2. A Linear function is called for s2, and likewise reshape is of the size c × h '× w', denoted as s _2.
And c, performing characteristic splicing on s _1 and s _2, and keeping the number of spliced channels unchanged after passing through one convolutional layer, wherein the result is recorded as conv _ cat _1. And performing characteristic splicing on the result out _ s1 after conv _ cat _1 convolution and s _0, and keeping the number of channels after splicing unchanged as c after passing through one convolution layer, wherein the result is recorded as conv _ cat _2.
And (3) acting the result out _ s2 after conv _ cat _2 convolution on the 2 convolution layers, calling a Linear function to obtain a vector s with dimension of Nx 1, namely a code word after the picture type data p is compressed and coded. Since the codeword s size nx1 is much smaller than the picture p original size c × h × w, as:
i.e. the size of the codeword s is only 1/64 of the original size of the picture p, the storage load and transmission bandwidth required for the picture type data can be greatly reduced.
Example 5
In this embodiment, referring to embodiment 1, in S7, the reconstruction processing of the image reconstruction module includes the following steps:
s71, inputting the code words into a convolution layer for convolution after linear transformation and matrix transformation to obtain a convolution result;
s72, after the characteristics of the primary convolution result are extracted, linear transformation and matrix transformation are carried out again to obtain a primary transformation result;
s73, after feature extraction is carried out on the primary convolution result twice, linear transformation and matrix transformation are carried out again to obtain a secondary transformation result;
s74, performing three-time feature extraction on the primary convolution result, and performing linear transformation and matrix transformation again to obtain a three-time transformation result;
s75, splicing the secondary linear transformation result and the tertiary linear transformation result, and performing convolution and twice convolution on the spliced result to obtain a secondary convolution result;
and S76, splicing the primary convolution result and the secondary convolution result, and convolving the spliced result to obtain a reconstructed picture with the same dimensionality as the picture type data.
The obtained codeword s is converted by a Linear function (Linear transformation function) and reshape is c × h × w, and the output result is recorded as i. And (3) acting i on a convolution layer, inputting the convolution result into a first UniFormer unit, and recording the output result of the UniFormer as i0.
Calling a Linear function for i0, and resetting the processing result to be c × h × w, which is recorded as i _0. And (3) acting i0 on a second UniFormer unit, circulating the unit twice to achieve a better picture reconstruction effect, and recording an output result as i1. Calling a Linear function for i1, and resetting the processing result to be c × h × w, which is denoted as i _1.
I1 is applied to the third UniFormer unit, which is also cycled twice, and the output is denoted as i2. A Linear function is called for i2, and the processing result reshape is c × h × w and is denoted as i _2. And c, performing characteristic splicing on i _1 and i _2, and keeping the number of spliced channels unchanged after passing through a convolution layer, wherein the result is recorded as conv _ cat _ 1'.
And performing characteristic splicing on the result out _ i1 after convoluting conv _ cat _1 'and i _0, and keeping the number of channels after splicing unchanged as c after passing through a convolution layer, wherein the result is recorded as conv _ cat _ 2'. The result out _ i2 after convoluting conv _ cat _ 2' acts on the 2 convolution layers to obtain a reconstructed picture with the same dimension as the original picture p
Example 6
In this embodiment, the present embodiment is described with reference to embodiment 1, and in S7, a cost function of the picture self-coding network model is a mean square error between the reconstructed picture and the original picture type data.
Designing cost function of the whole UniFormer-based AutoEncoder model as picture type data reconstructed by a decoder
Mean square error with the original picture type data p, i.e. a cost function, of
Wherein C is the total number of training samples,‖·‖ 2 is the euclidean norm.
The parameters of the encoder and decoder (mainly including weights, offsets and convolution kernels) can be jointly trained by using a Ranger optimization algorithm and an end-to-end learning mode, so that the cost function is minimum. The trained UniFormer-based AutoEncoder model can be used for compression coding and decoding reconstruction of subsequent picture type data.
Example 7
In the data security encryption method disclosed in this embodiment, in S2, the feature extraction layer includes a dynamic position code, a multi-head relationship aggregator, and a feed-forward neural network.
The feature extraction layer uniform unit is composed of a Dynamic Position Encoding (DPE), a Multi-Head relationship Aggregator (MHRA) and a Feed Forward neural Network (FFN). The feed-forward neural network, i.e. the input layer, hidden layer or output layer neural structures which are parallel to each other, flows information from front to back.
Example 8
The embodiment is described with reference to embodiment 1, and the data security encryption device disclosed in this embodiment includes a text encryption module, a picture compression module, a picture encryption module, a decryption module, and a picture reconstruction module;
the text encryption module is used for encrypting the original text type data so as to obtain text ciphertext data;
the picture compression module is used for pre-compressing the original picture type data so as to obtain picture compression codes;
the picture encryption module is used for encrypting the picture compression codes so as to obtain picture ciphertext data;
the decryption module is used for decrypting the text ciphertext data or the picture ciphertext data which need to be applied to a downstream task according to user requirements to obtain decrypted text type data and decrypted picture compression codes;
the picture reconstruction module is used for reconstructing and restoring the picture compression codes so as to obtain reconstructed picture type data.
Claims (8)
1. A data security encryption method is characterized by comprising the following steps:
s1, dividing all data to be processed into original text type data and original picture type data;
s2, encrypting original text type data in data to be encrypted by adopting a text encryption module to obtain text ciphertext data;
s3, constructing a picture self-coding network model, and pre-compressing original picture type data to be encrypted by adopting a picture compression module to obtain picture compression codes; the picture self-coding network model comprises an encoder and a decoder, wherein the encoder compresses and encodes picture type data to form a code word;
s4, encrypting the picture compression codes by adopting a picture encryption module to obtain picture ciphertext data;
s5, transmitting or storing the text ciphertext data and the picture ciphertext data according to user requirements;
s6, decrypting the text ciphertext data or the picture ciphertext data which need to be applied to the downstream task by adopting a decryption module to obtain decrypted text type data and decrypted picture compression codes;
and S7, reconstructing and restoring the decrypted picture compression codes by adopting a picture reconstruction module, and reconstructing the code words by the decoder to obtain reconstructed picture type data.
2. The data security encryption method of claim 1, wherein in S2 and S4, the encryption processing methods adopted by the text encryption module and the picture encryption module include a symmetric encryption algorithm and/or an asymmetric encryption algorithm.
3. A data security encryption method according to claim 1, wherein in S3, said encoder section includes an input layer, a convolutional layer and a feature extraction layer, and said decoder includes a convolutional layer, a feature extraction layer and an output layer.
4. A secure data encryption method according to claim 1 or 3, wherein in S3, the pre-compression process adopted by the picture compression module comprises the following steps:
s31, taking the original picture type data as the input of the encoder part, and defining the depth, the width and the height of the original picture type data;
s32, passing the defined picture type data through a convolution layer, performing linear transformation and coding twice on the result after convolution, and splicing the result after the linear transformation twice and the result after the linear transformation once to obtain a convolution splicing result;
and S33, carrying out secondary convolution on the primary convolution splicing result through a convolution layer to enable the depth of the primary convolution splicing result to be unchanged, splicing the result after the secondary convolution and the result of the primary convolution and carrying out linear transformation to obtain a code word with the unchanged depth and the width and the height reduced to 1/8 of the original picture type data.
5. The data security encryption method according to claim 1, wherein in S7, the reconstruction process of the picture reconstruction module comprises the following steps:
s71, inputting the code words into a convolution layer after linear transformation and matrix transformation to carry out primary convolution to obtain a primary convolution result;
s72, after the characteristics of the primary convolution result are extracted, linear transformation and matrix transformation are carried out again to obtain a primary transformation result;
s73, after feature extraction is carried out on the primary convolution result twice, linear transformation and matrix transformation are carried out again to obtain a secondary transformation result;
s74, performing three-time feature extraction on the primary convolution result, and performing linear transformation and matrix transformation again to obtain a three-time transformation result;
s75, splicing the secondary linear transformation result and the tertiary linear transformation result, and performing convolution and twice convolution on the spliced result to obtain a secondary convolution result;
and S76, splicing the primary convolution result and the secondary convolution result, and convolving the spliced result to obtain a reconstructed picture with the same dimensionality as the picture type data.
6. A data security encryption method according to claim 1 or 5, characterized in that in S7, the cost function of the picture self-coding network model is the mean square error of the reconstructed picture and the original picture type data.
7. A method for secure encryption of data according to any of claims 1-3, wherein in S2, the feature extraction layer comprises dynamic position coding, a multi-head relationship aggregator and a feed-forward neural network.
8. A data security encryption device is characterized by comprising a text encryption module, a picture compression module, a picture encryption module, a decryption module and a picture reconstruction module;
the text encryption module is used for encrypting the original text type data so as to obtain text ciphertext data;
the picture compression module is used for pre-compressing the original picture type data so as to obtain picture compression codes;
the picture encryption module is used for encrypting the picture compression codes so as to obtain picture ciphertext data;
the decryption module is used for decrypting the text ciphertext data or the picture ciphertext data which need to be applied to a downstream task according to user requirements to obtain decrypted text type data and decrypted picture compression codes;
the picture reconstruction module is used for reconstructing and restoring the picture compression codes so as to obtain reconstructed picture type data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210801788.0A CN115190217B (en) | 2022-07-07 | 2022-07-07 | Data security encryption method and device integrating self-coding network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210801788.0A CN115190217B (en) | 2022-07-07 | 2022-07-07 | Data security encryption method and device integrating self-coding network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115190217A true CN115190217A (en) | 2022-10-14 |
| CN115190217B CN115190217B (en) | 2024-03-26 |
Family
ID=83517357
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210801788.0A Active CN115190217B (en) | 2022-07-07 | 2022-07-07 | Data security encryption method and device integrating self-coding network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115190217B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1745580A (en) * | 2003-03-07 | 2006-03-08 | 松下电器产业株式会社 | Encryption device, decryption device, and data reproduction device |
| CN101894115A (en) * | 2009-05-18 | 2010-11-24 | 北京大学 | Image data processing method of electronic document and device thereof |
| CN104506793A (en) * | 2015-01-16 | 2015-04-08 | 东北大学 | Safe video conference system and method based on internet |
| CN106791827A (en) * | 2016-12-13 | 2017-05-31 | 北京黎阳之光科技有限公司 | A kind of Novel Image Compression Coding device |
| US20170237715A1 (en) * | 2014-08-07 | 2017-08-17 | Gurulogic Microsystems Oy | Encoder, decoder and method |
| US20170279604A1 (en) * | 2014-09-19 | 2017-09-28 | Gurulogic Microsystems Oy | Encoder, decoder and methods employing partial data encryption |
| CN107423309A (en) * | 2016-06-01 | 2017-12-01 | 国家计算机网络与信息安全管理中心 | Magnanimity internet similar pictures detecting system and method based on fuzzy hash algorithm |
| CN108664953A (en) * | 2018-05-23 | 2018-10-16 | 清华大学 | A kind of image characteristic extracting method based on convolution self-encoding encoder model |
| CN109559358A (en) * | 2018-10-22 | 2019-04-02 | 天津大学 | It is a kind of that the method for sampling is risen from the image pattern of coding based on convolution |
| CN109889340A (en) * | 2019-03-19 | 2019-06-14 | 北京信息科技大学 | A kind of adaptive cipher key distribution mechanisms based on QKD network |
| CN111160095A (en) * | 2019-11-26 | 2020-05-15 | 华东师范大学 | Unbiased face feature extraction and classification method and system based on depth self-encoder network |
| CN111401272A (en) * | 2020-03-19 | 2020-07-10 | 支付宝(杭州)信息技术有限公司 | Face feature extraction method, device and equipment |
| CN113658032A (en) * | 2021-10-20 | 2021-11-16 | 广东迅维信息产业股份有限公司 | Image watermark encryption and decryption method and system based on deep learning and image processing |
-
2022
- 2022-07-07 CN CN202210801788.0A patent/CN115190217B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1745580A (en) * | 2003-03-07 | 2006-03-08 | 松下电器产业株式会社 | Encryption device, decryption device, and data reproduction device |
| CN101894115A (en) * | 2009-05-18 | 2010-11-24 | 北京大学 | Image data processing method of electronic document and device thereof |
| US20170237715A1 (en) * | 2014-08-07 | 2017-08-17 | Gurulogic Microsystems Oy | Encoder, decoder and method |
| US20170279604A1 (en) * | 2014-09-19 | 2017-09-28 | Gurulogic Microsystems Oy | Encoder, decoder and methods employing partial data encryption |
| CN104506793A (en) * | 2015-01-16 | 2015-04-08 | 东北大学 | Safe video conference system and method based on internet |
| CN107423309A (en) * | 2016-06-01 | 2017-12-01 | 国家计算机网络与信息安全管理中心 | Magnanimity internet similar pictures detecting system and method based on fuzzy hash algorithm |
| CN106791827A (en) * | 2016-12-13 | 2017-05-31 | 北京黎阳之光科技有限公司 | A kind of Novel Image Compression Coding device |
| CN108664953A (en) * | 2018-05-23 | 2018-10-16 | 清华大学 | A kind of image characteristic extracting method based on convolution self-encoding encoder model |
| CN109559358A (en) * | 2018-10-22 | 2019-04-02 | 天津大学 | It is a kind of that the method for sampling is risen from the image pattern of coding based on convolution |
| CN109889340A (en) * | 2019-03-19 | 2019-06-14 | 北京信息科技大学 | A kind of adaptive cipher key distribution mechanisms based on QKD network |
| CN111160095A (en) * | 2019-11-26 | 2020-05-15 | 华东师范大学 | Unbiased face feature extraction and classification method and system based on depth self-encoder network |
| CN111401272A (en) * | 2020-03-19 | 2020-07-10 | 支付宝(杭州)信息技术有限公司 | Face feature extraction method, device and equipment |
| CN113658032A (en) * | 2021-10-20 | 2021-11-16 | 广东迅维信息产业股份有限公司 | Image watermark encryption and decryption method and system based on deep learning and image processing |
Non-Patent Citations (2)
| Title |
|---|
| GUOXIANG TONG: ""A Residual Chaotic System for Image Security and Digital Video Watermarking"", IEEE ACCESS * |
| 季秋菊: ""基于BOW的密文图像检索研究"", 中国优秀硕士学位论文全文数据库 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115190217B (en) | 2024-03-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Qian et al. | New framework of reversible data hiding in encrypted JPEG bitstreams | |
| Qian et al. | Reversible data hiding in encrypted JPEG bitstream | |
| CN102891999B (en) | Combined image compression/encryption method based on compressed sensing | |
| JP6608436B2 (en) | Encoder, decoder and method using partial data encryption | |
| Zhou et al. | Scalable compression of stream cipher encrypted images through context-adaptive sampling | |
| CN105426709A (en) | JPEG image information hiding based private information communication method and system | |
| KR20030043620A (en) | Encoding/decoding method and apparatus for key value of coordinate interpolator node | |
| Manohar et al. | Data encryption & decryption using steganography | |
| CN112911290B (en) | Ciphertext image reversible data hiding method based on predictive difference compression coding | |
| TW200830732A (en) | Image encoding method and decoding method, source encoding method and decoding method, apparatuses therefor, programs therefor, and storage media for storing the programs | |
| CN104869432B (en) | A kind of image multiresolution compression encryption and classification decryption and display methods | |
| CN112491832B (en) | File transmission method and device | |
| CN100456668C (en) | Information processing method, information processing apparatus, program, and storage medium | |
| Suguna et al. | Secure image communication through adaptive deer hunting optimization based vector quantization coding of perceptually encrypted images | |
| CN113660386B (en) | Color image encryption compression and super-resolution reconstruction system and method | |
| Merhav | Perfectly secure encryption of individual sequences | |
| Singh et al. | An improved encryption–Compression-based algorithm for securing digital images | |
| CN104077541A (en) | Selective encryption method and device based on multimedia data | |
| Cao et al. | Separable Reversible Data Hiding in Encrypted VQ‐Encoded Images | |
| CN115190217B (en) | Data security encryption method and device integrating self-coding network | |
| Kumar et al. | An efficient compression of encrypted images using WDR coding | |
| Im et al. | FaceBERT: Face De-Identification Using VQGAN and BERT | |
| KR101330071B1 (en) | Security apparatus and method for protecting numeric data in database | |
| Chen et al. | On the design of a two-decoding-option image secret sharing scheme | |
| CN115150055A (en) | Privacy protection ridge regression method based on homomorphic encryption |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |